It is well known to use speech coding in communication systems to reduce the bandwidth required for the transmission of speech. In wireless communication systems, and more specifically cellular radiotelephone systems, speech coding rates less than 16 kbps are generally used The achievable quality of these coders is somewhat less than "toll quality" which is basically that level of quality given by typical land-line telephone systems where speech is coded at 64 kbps. Generally, as speech coding rates decrease, the level of quality correspondingly decreases.
In wireless communication systems, the measure of quality of a particular type/rate of speech coder is given by a mean opinion score (MOS). The MOS is a subjective scoring system, having a scoring range between 1-5 or between poor to excellent. A listener rates the particular type/rate coder between the ranges when compared to other types/rates of coders. The higher the rating, the better the speech sounded to the listener.
In cellular radiotelephone systems, and more particularly digital cellular radiotelephones systems, tandem speech coding scenarios will exist at certain times. In tandem speech coding scenarios, a speech input signal is not coded only once, but may be coded twice or more. A common example is when a cellular mobile user desires to leave or retrieve a message on a voice mail system. Not only does the cellular system code the speech input, but the voice mail system may likewise code the speech input signal according to the same or different algorithm. In an example of such a tandem speech coding scenario, where a tandem coding of two vector sum-excited linear predictive (VSELP) speech coders is utilized, the MOS score is reduced from 3.85 for single coding to 3.13 for tandem coding. Thus a need a exists for a method and apparatus for coding speech which reduces excessive degradation in tandem speech coding scenarios.